The pulmonary artery catheter: a solution still looking for a problem

Abstract

In 1970, Jeremy Swan, William Ganz, and colleagues published their seminal paper ‘‘Catheterization of the heart in man with use of flow-directed balloon-tipped catheter’’. Few devices in our specialty have generated more interest or passion as the pulmonary artery catheter (PAC). At its core, the PAC measures right-sided cardiac pressures, pulmonary capillary wedge pressure, cardiac output, and mixed venous oxygen saturation. Given the ability to guide therapy in near real-time to subtle perturbations in physiology, the PAC became synonymous with critical care medicine. Clinicians have believed that using the PAC to ‘‘correct’’ abnormal physiology results in better clinical outcomes. Nevertheless, there are fundamental flaws that underpin this belief. First, the PAC is simply a monitor, it provides no therapy by itself. Any benefit is crucially dependent on accurate and reproduceable interpretation by the clinician. However, clinicians often do not agree on the interpretation of PAC values and can act upon the same physiologic data with diametrically opposed interventions. Second, the PAC carries attributable risk with both misinterpretation of data and procedural complications. Finally, studies have consistently failed to show any benefit to patients in whom a PAC is inserted. In 1991, a randomized trial of PAC use in critically ill patients in Ontario was stopped early as only 33 of 148 eligible patients were randomized. Fifty-two eligible patients were excluded as the attending physician felt it was unethical to withhold a PAC. In 1996, Connors et al. published a cohort study of 5,735 critically ill patients showing the PAC use was associated with increased mortality and resource utilization.This served as a call to action to conduct more high-quality research on the use and indications of PAC. In the early 2000’s, several randomized trials failed to show any superiority of PAC use over standard care in different populations, including high-risk surgical patients, general critical care patients, and patients with acute respiratory distress syndrome. This lack of benefit persisted even when the interventions were standardized, thereby obviating the need for physicians to correctly interpret the data. In 2005, the multicentre Evaluation Study of Congestive Heart Failure and Pulmonary Artery Catheterization Effectiveness (ESCAPE) trial showed that when added to clinical assessment in patients with severe symptomatic and recurrent heart failure, the PAC afforded no improvement in any clinical outcomes, while doubling adverse events. Interestingly, sites participating in ESCAPE kept a registry of patients in whom the clinician felt that a PAC was indicated, and thus were not randomized. These 439 patients who received a PAC outside of the trial were more severely ill (lower blood pressure, worse renal function, and higher use of inotropes) and had higher mortality and longer hospitalizations than patients enrolled in the ESCAPE trial. Thus, even clinicians in participating sites, where there was presumably S. Thiara, MD Division of Critical Care Medicine, Department of Medicine, The University of British Columbia, ICU Room 2438, JPP2, 899 West 12th Avenue, Vancouver, BC V5Z 1M9, Canada